2009ApJ...701.1519R

The abundance of beryllium in the oldest, most metal-poor stars acts as a probe of early star formation and Galactic chemical evolution. We have analyzed high-resolution, high signal-to-noise ratio Keck/HIRES spectra of 24 stars with [Fe/H] from -2.3 to -3.5 in order to determine the history of Be abundance and explore the possibility of a Be plateau. We have determined stellar parameters of our sample spectroscopically, using equivalent widths of Fe I, Ti I, and Ti II lines. We have determined O abundances from three OH features which occur in the same spectral region; this region is relatively uncrowded and has a well determined continuum in these very/extremely metal-poor stars. We have supplemented this sample with reanalyzed spectra of 25 stars from previous observations so that our total sample ranges in [Fe/H] from -0.5 to -3.5. Our results indicate that the relationship between Be and [Fe/H] continues to lower metallicities with a slope of 0.92±0.04. Although there is no indication of a plateau with constant Be abundance, the four lowest metallicity stars (below [Fe/H] of -3.0) do show a Be enhancement relative to Fe at the 1σ level. A single relationship between Be and [O/H] has a slope of 1.21±0.08, but there is also a good fit with two slopes: 1.59 above [O/H] = -1.6 and 0.74 for stars with [O/H] below -1.6. This change in slope could result from a change in the dominant production mechanism for Be. In the era of the formation of the more metal-poor stars, Be would be formed by acceleration of CNO atoms in the vicinity of SN II and in later times by high-energy cosmic-rays bombarding CNO in the ambient interstellar gas. We find an excellent correlation between [Fe/H] and [O/H] and show that [O/Fe] is near +1.0 at [Fe/H] = -3.5 declining to 0 at [Fe/H] = 0.